Background:
- Having my cluster on UTC means that Debian is choosing the .nl time servers.
- In Check_MK, I see that these checks sometimes don't happen for up to two hours (alerts at 35 minutes without NTP check).
- Rather than change the factory PVE settings, I prefer to either adjust monitoring parameters or find a "built-in" PVE solution.
- Drift is between .2 and .5ms but during long periods without NTP contact, it has risen as high as 1.5 - 2.0 ms.
- If I change my tzdata in PVE to my local time, will this affect corosync/ceph? Will they detect a problem and choke?
- Is there another reason why chrony is failing to NTP sync regularly with the default time servers?